comptime compiler-API: Phase 1 foundation + Phase 2.1 weld plan

Introduce the welded comptime `compiler` library (`#library "compiler"` +
`abi(.zig) extern compiler`), per design/comptime-compiler-api.md, and unify
`callconv(...)` into the new `abi(...)` annotation.

abi(...) replaces callconv(...):
- New ABI enum { default, c, zig, pure }; `abi(.c|.zig|.pure)` parses in the
  postfix slot before extern/export (and standalone). `kw_callconv` -> `kw_abi`.
- Migrated 52 sx files, the call-convention-mismatch diagnostic, and docs
  (readme/specs) from `callconv(.c)` to `abi(.c)`.

Phase 1 — welded compiler library (parse -> registry -> validation -> bridge):
- `abi(.zig) extern compiler` parses on fn decls (carries abi/extern_lib) and
  struct decls (StructDecl.abi/extern_lib).
- `#library "compiler"` is the comptime-only internal surface — never dlopen'd.
- src/ir/compiler_lib.zig: the binding registry (the safety boundary). `Field`
  welded to StructInfo.Field with layout baked from the real Zig type
  (@offsetOf/@sizeOf); `findType`/`findFn`. Welded structs are layout-validated
  at registration (field set + total size) as a header checked against the impl.
- Host-call bridge: a `fn abi(.zig) extern compiler` dispatches under the
  comptime interp to its registered Zig handler (intern/text_of round-trip),
  never dlsym. IR Function.compiler_welded; validated in declareFunction.
- Comptime-only enforcement: a runtime call to a welded fn is a clean
  build-gating error (emitCall), not an undefined-symbol link failure.

Phase 2.1 — byte-layout weld foundation:
- Decision: full byte-layout weld (sx struct laid out byte-identically to the
  bound Zig type). Registered StructInfo (first non-natural / Zig-reordered
  layout). `computeWeldPlan` — pure offset-ordered element plan + padding +
  sx-field->LLVM-element remap; unit-tested. Emit/interp wiring is the next
  sub-step (2.2+, see current/CHECKPOINT-COMPILER-API.md).

Examples: 0625/0626 (welded struct + fn round-trip), 1183/1184/1185
(layout-mismatch, unexported-fn, runtime-call diagnostics).

library/modules/std/thread.sx

@@ -2,7 +2,7 @@
 // pthreads (PLAN-HTTPZ S6).
 //
 // THE RE-ENTRY CONTRACT (pinned by examples/1636): a thread entry is a
-// `callconv(.c)` function — it has NO implicit context — and enters
+// `abi(.c)` function — it has NO implicit context — and enters
 // the sx world by fabricating one: `push Context.{ allocator = xx gpa }`
 // around the default-conv code it runs. Pool workers do exactly that,
 // each with its own malloc-backed GPA, so tasks allocate freely and
@@ -26,7 +26,7 @@
 
 tlib :: #library "c";
 
-pthread_create  :: (thread: *usize, attr: *void, start: (*void) -> *void callconv(.c), arg: *void) -> i32 extern tlib;
+pthread_create  :: (thread: *usize, attr: *void, start: (*void) -> *void abi(.c), arg: *void) -> i32 extern tlib;
 pthread_join    :: (thread: usize, retval: **void) -> i32 extern tlib;
 pthread_detach  :: (thread: usize) -> i32 extern tlib;
 
@@ -104,9 +104,9 @@ Cond :: struct {
 Thread :: struct {
     handle: usize = 0;
 
-    // `entry` is the C->sx boundary: callconv(.c), fabricates its own
+    // `entry` is the C->sx boundary: abi(.c), fabricates its own
     // Context before touching default-conv sx code (examples/1636).
-    spawn :: (entry: (*void) -> *void callconv(.c), arg: *void) -> (Thread, !ThreadErr) {
+    spawn :: (entry: (*void) -> *void abi(.c), arg: *void) -> (Thread, !ThreadErr) {
         t : Thread = .{};
         if pthread_create(@t.handle, null, entry, arg) != 0 { raise error.Spawn; }
         return t;
@@ -200,7 +200,7 @@ Pool :: struct {
 
 // The worker loop: C entry, own fabricated Context, then
 // pop-task/run-task until stop with an empty queue.
-pool_worker :: (arg: *void) -> *void callconv(.c) {
+pool_worker :: (arg: *void) -> *void abi(.c) {
     p : *Pool = xx arg;
     gpa := GPA.init();
     push Context.{ allocator = xx gpa } {